This invention relates to sensing proximity of a target using an electronic sensor, and more particularly, using a sensor that responds to near-field electromagnetic effects to sense the position of a small object. This invention also relates to sensing the position of a gear by determining when a signal voltage is at some predetermined point between two fixed extremes and furnishes a change of state when the signal voltage reaches a predetermined threshold value between these extremes.
Numerous proximity sensors are known in the art that react to various parameters of the target: These electromagnetic proximity sensor devices, such as Hall effect devices, Wiegand effect devices, eddy-current killed oscillators, etc., have the general disadvantage that they require that the target be constructed of a ferrous or magnetic material and that the device be located in very close proximity to the target. Further, these devices cannot distinguish between non-target objects that move in close proximity to the target and the target itself. Other electromagnetic proximity sensor devices such as microwave proximity sensors use radar techniques, such as the Doppler effect, to sense large targets at long distances from the sensor. Prior art electromagnetic devices are not designed to sense small objects at close distances and are both bulky and expensive.
The most common use of a proximity sensing circuit is tracking the rotational movement of a gear. As each gear tooth passes across the front of the proximity sensing circuit, a changing signal is generated. The changing signal voltage is highest when the sensor is directly in front of the face of a gear tooth and lowest when facing the valley between gear teeth. Allowing for small tolerances, the maximum voltage will be the same for each tooth, and the minimum voltage will be the same for each valley. Therefore, the signal will always fall between these two values.
When the gear is rotating at constant velocity, the signal will resemble a sine wave. The proximity sensor has a set area in which objects influence its output. If the gear tooth or valley is larger than the area of influence of the sensor, the signal resulting from its movement will tend to remain at a maximum or minimum value for some period of time. Also, since signal waveform at these two extremes is relatively flat, any electromagnetic noise from the environment surrounding the sensor or from the proximity sensor components themselves will be superimposed on the waveform.
The result is that the maximum and minimum vales are not suitable as trigger points to reliably cause the output state of the signal conditioner to change. This is because each peak of the noise voltage may cause the signal conditioner to trigger. Many of these peak voltages may be present on the signal, causing any number of state changes for a single gear tooth. The best point at which to trigger is some point midway between the maximum and minimum. This is because the signal waveform at this time has a steep slope. This ensures that the voltage at any one time is probably at the threshold value only once during the transition from maximum to minimum. Any noise riding on the signal will change in level rapidly as a result of its riding on the steep slope. Should any noise cause false triggering during this time, the error, if any, (expressed in degrees of rotation) will be very small due to the relatively small amount of time that the signal remains at the trigger point. This is the point at which a gear tooth is halfway across the front of the sensor.
Another common task for the proximity sensor is to measure the opening and closing of a valve. The sensor is normally placed so the valve approaches the sensor when opening and goes away from the sensor when closing. For this application, the signal conditioning circuitry should change state when the flow begins, and again when the flow ends. Due to inertial forces, the valve does not immediately become fully open, but begins to rise relatively slowly, reaches its peak, falls to its rest position, then may bounce once or a few times. The optimum threshold point for this application is just above the signal value obtained when the valve is at rest. However, noise may again be present to cause false triggering, and second, the bounce signal may cause the signal conditioner to change states each time the valve bounces. The signal conditioner can be prevented from being triggered by the noise and bounce signals by moving the threshold just above the opening point. This does not pose a problem in most applications, as little flow occurs until the valve is an appreciable distance off its seat. The optimum threshold in this case is approximately 2 to 10 percent of the total movement of the valve.
To convert the signal to a square-wave pulse the signal is usually compared to a DC threshold voltage in a comparator. When the signal is higher in voltage than the DC threshold, the comparator usually delivers a positive going pulse. When the signal falls below the threshold, the comparator changes back to the off state with an output near zero volts.
By the proper selection of the threshold voltage, the comparator can be made to switch when the target is at some percentage of distance between its minimum and maximum excursions. The key is knowing the precise signal voltage at this point.
The problem lies in the fact that due to manufacturing tolerances, various target sizes and movements, and varying installations, the signal can range over a wide value of voltages. Installations may require differing distances from the sensor to the target, and mechanical tolerances may mean that identical targets in similar mechanisms move different amounts. Sensor manufacturing tolerances may deliver different gains and internal reference voltages that can cause different signal voltages from each proximity sensor at both the minimum and maximum excursions of identical targets. Any stationary targets in the area of influence of the proximity sensor will also add to the signal voltage.
Because of these factors, a signal minimum and maximum may range from 4 to 5 volts, another from 3 to 4, another from 1 to 1.5, etc. The task for the signal conditioner is to determine the difference between the maximum and the minimum signal voltages and to use some percentage of this difference to generate a threshold voltage to cause the comparator to change state at the appropriate time.
Historically, there have been two general methods for solving this problem. The analog approach attempts to solve this problem by passing the signal through a capacitor to return any offset back to zero volts. (This means that a signal of 4 to 5 volts becomes 0 to 1 volt, a signal of 3 to 5 volts becomes 0 to 2, etc.)
The problem with feeding the signal through a capacitor is that a capacitor effectively blocks any signal from the sensor that is very low in frequency (as is the case with slow moving targets). This means that the analog method cannot be used for zero speed detection. Below a certain rate of movement, the sensor will not be able to tell that the target is moving because the signal is so close to a steady DC voltage that it will not pass through the capacitor.
A digital approach has been pursued wherein the signal from the target has been converted to a digital number, stored in memory, then reconverted into an analog voltage to use as the threshold. The problems with this approach are: It requires a very large number of components to be realized, thus is complex, relatively expensive, and is large in area. The second problem is that it suffers from two conversion inaccuracies: when the signal is converted from analog to digital, and again when it is converted back to analog form. The solution of either one of these two problems causes an increase in the problems caused by the other condition. The overall accuracy can be increased only by increasing the component count exponentially, and the component count can only be reduced by decreasing the accuracy.